Responsible AI has been a popular topic in academic and industry settings with the advent of conversational AI based on generative models in the last two years. Despite the growing scientific research in this field, what should be considered when designing for responsibility in conversational AI interactions is still being investigated. We will discuss responsible AI in public and private cooperative settings and human values identified as essential to designing responsible conversational systems Our research projects have investigated how to foster trust, accountability, transparency, fairness, and acceptance of conversational user interfaces deployed in natural settings with diverse audiences. I will describe unveiling practices extracted from empirical research with content and tech curators in training conversational machines in work settings. Furthermore, we will discuss how bias emerged as a criterion when interacting with collaborative AI exhibitions in museum settings and our recent work in elucidating values such as creditworthiness with micro businesswomen in underrepresented communities using conversational systems. This talk can serve as the basis for further discussions during the conference on promoting social impact and mitigating harm when designing conversational systems with responsibility.

Recently, there is a growing trend of using generative AI systems and tools for fostering and protecting online collaborative communities. Yet, existing AI tools may introduce new risks and even harm to diverse communities’ online safety. How to better maximize the novel opportunities of AI and mitigate its emerging risks and harm for our future online safety is a critically needed discussion for the HCI community. Featuring experts from both industry and academia, the goal for this panel is to promote interdisciplinary, community-wide discussions and collective reflections on important questions and considerations at the unique intersection of AI and online communities, including but not limited to: how the design of AI systems may discourage existing online harm but also invite new online harm in various online spaces; how different populations, cultures, and communities may perceive and experience AI’s new roles for their online safety; and what new strategies, principles, and directions can be envisioned and identified to better design future AI technologies to protect rather than harm various online communities.

Despite resurgent interest in remote work technologies, many forms of specialised work remain inaccessible. Participation in remote activities often requires more than video-conferencing. Telepresence technologies, using robotics and mixed reality, can fill that gap by providing more affordances (movement and embodied expressions). However, we still need to know what kinds of affordances are needed and how they should be implemented. This workshop focuses on telepresence technologies for mediating specialised tasks and extending expertise reach, such as teaching technical skills or providing healthcare assistance. We will invite experts from specific fields (e.g., education, healthcare, industry) to guide participants through use cases, and use interactive exercises to identify what kinds of actions and interaction modalities are important in different scenarios. The aim of the workshop is to explore this hands-on, needs-driven design approach, initiate critical discussions into the design of telepresence technologies and foster future work driven by a focus on understanding practical user needs.

This study explores robotic telepresence in classrooms, from the perspectives of both the remote students (operators) and their in-person classmates. Analyzing a subset of field study data from 35 participants (22 operators, 13 classmates), we identified differing priorities: operators of the robot emphasize the need for improved functional aspects like camera control, while co-present classmates value social presence, interaction and identification, advocating for features like personalization and effective communication channels. These insights underline the need for telepresence robots that balance functional efficacy with social integration to provide a holistic educational experience. This work, though incomplete, enriches Group and CSCW literature on telepresence robots to facilitate collaborative interactions by highlighting the distinct needs of remote and in-person students, proposing design recommendations for telepresence robots that are technically proficient and socially responsive.

Live coding is a performance art in which media, commonly music and video, are generated and manipulated with open-source tools. While live coding initially centered upon local meet-ups, many individuals and loosely organized groups now learn, collaborate, and share work online. This expansion of interest may offer new, informal opportunities to learn computer programming, but little is known about how releveant online forums operate. To better understand this online space, we gathered and analyzed content from 414 threads in two popular live coding forums, Tidal Club and in_thread. We found that while both communities include discussions about using and extending core programming tools, significantly more posts on Tidal Club pertain to installation difficulties, while in_thread possesses a wider culture of sharing creative artifacts. We reflect on these findings and offer suggestions for future work in understanding the motivations behind why live coders post code, how they receive help, and the extent to which informal learning in live coding forums aligns with formal computer programming and music teaching.

Augmented Reality (AR) technology holds significant potential to transform educational settings and improve learning, especially in complex disciplines such as chemistry, where abstract concepts and reactions often pose challenges for students. Despite this potential, the classical educational settings contains and implies barriers related to the practice-oriented design and user-friendly interactions of AR. This work present “Virtual Chemist”, an AR-based system that was designed together with overall 28 students and teachers in secondary-level chemistry education. We present here the outcomes of a participatory design case study, in which we conducted requirements, co-designed the system as well as observed their appropriation and interaction with the participants. Our results indicate that the system enhanced student motivation and engagement; students reported that the application was enjoyable and user-friendly, particularly valuing its capacity to elucidate complex chemical bonds and reactions through three-dimensional visualizations. The prototype and approach presented here will serve to discuss and reflect future research activities, methodological concepts, and experiences in the field of HCI and educational AR.

Assistive technologies for people with visual impairments (PVI) have made significant advancements, particularly with the integration of artificial intelligence (AI) and real-time sensor technologies. However, current solutions often require PVI to switch between multiple apps and tools for tasks like image recognition, navigation, and obstacle detection, which can hinder a seamless and efficient user experience. In this paper, we present NaviGPT, a high-fidelity prototype that integrates LiDAR-based obstacle detection, vibration feedback, and large language model (LLM) responses to provide a comprehensive and real-time navigation aid for PVI. Unlike existing applications such as Be My AI and Seeing AI, NaviGPT combines image recognition and contextual navigation guidance into a single system, offering continuous feedback on the user’s surroundings without the need for app-switching. Meanwhile, NaviGPT compensates for the response delays of LLM by using location and sensor data, aiming to provide practical and efficient navigation support for PVI in dynamic environments.

In the context of programming education, programming clubs serve as an innovative non-formal environment that supplements traditional classroom instruction. However, the practices within these non-formal learning environments and their impact on participants remain unclear. Our study first employed observational study to understand the internal dynamics and formality of these programming clubs. Subsequently, we conducted semi-structured interviews (n = 7) to explore participants’ programming learning experiences in both formal and non-formal settings, and analyzed the impact of differences in learning formats. Ultimately, our research identified three types of activities in coding clubs and six dimensions for measuring the formality of learning. We analyzed how changes in these formality dimensions affected participants’ perceptions. This preliminary study provides insights that can enhance the effectiveness of these extracurricular non-formal programming learning activities.

Selfies have become a common and popular form of communication and self-representation in the digital age, particularly for the purpose of posting on social media. While the phenomenon has been widely studied, this paper seeks to examine motivations for selfie-taking specific to Instagram through in-depth interviews with 25 young adults. We identified six key themes behind why people take selfies: emphasizing physical appearance, expressing personal identity, conveying authenticity, documenting life experiences, fostering social connections, and seeking popularity. This study sheds light on how selfies serve multiple roles, from social engagement to personal reflection, and highlights how technology continues to shape identity.

Designing explainable and personalized AI systems to provide support to older adults aging in place requires an understanding of their motivations and expectations for the explanations. This poster presents our ongoing work in exploring explanation preferences within AI systems for older adults aging in place with their carepartners. We do so by leveraging the speculative and iterative benefits of the Research through Design (RtD) approach in HCI, and explore variations in explanation requirements for different users by understanding their needs, goals and motivations for the different sources of information within the home. We illustrate an example for employing a Research through Design inquiry for the design of AI applications, adopting speculative methods to probe into future possibilities of Explainable AI (XAI) using a human-centered design framework. Through a Speed Dating study and a Co-Design activity, we investigate different explanation types and scenarios and argue for a shift in the algorithmic focus of Explainable AI research toward user-centered requirements, positioning explanation as a collaborative process between AI systems and users.

This paper examines the impact of platform migration on the QAnon community during their transition from Twitter to alternative platforms like Parler and Dotwin following initial widespread account bans on Twitter. We examine how QAnon’s community dynamics—including user roles and activities—evolved amidst this migration by analyzing their activities on Twitter, Parler, and Dotwin in the months leading up to the Jan 6th Capitol attack. We assess user engagement and influence changes, categorizing users into five distinct roles: ’common users,’ ‘broadcasters,’ ’influentials,’ ’hidden influentials,’ and ’lurkers.’ Our findings challenge traditional linear migration models by revealing extremist communities’ complex, multi-platform engagement during early-stage migration. Our findings also suggest the importance of platform affordances in supporting various user roles in extremist communities. Our research also broadens the scope of existing literature by offering a longitudinal cross-platform analysis, providing new insights into the evolving dynamics of extremist online activity.

E-cigarettes are the most popular tobacco product among youth, yet existing prevention and cessation interventions are primarily adult-centered and adapted from general tobacco use, with limited evidence for their effectiveness in addressing youth e-cigarette use. This paper explores technology-based preventative measures targeting youth e-cigarette use by systematically reviewing 90 articles. Our findings highlight a significant research gap in youth-specific e-cigarette interventions, despite the increasing prevalence and associated risks. The most-studied vaping prevention method was based on virtual reality games, showing positive outcomes in terms of youth engagement. Overall, we emphasize the critical need for vaping-specific, youth-centered research and design to meet their unique needs in combating their e-cigarette use.

Social media has transformed how fashion trends are influenced, with fashion influencers playing a crucial role in shaping public perception and purchasing decisions. To better understand and visualize the influence of these key figures, we present FITViz, an interactive visualization framework designed to analyze the interactions between social media influencers within the fashion domain. FITViz leverages data from Twitter to uncover patterns of communication, network effects, and the spread of influence among influencers, brands, media, and niche communities. Through a user study, we demonstrate the tool’s effectiveness in aiding non-technical professionals in exploring influencer interactions and gaining insights into fashion trends. Our findings highlight the utility of FITViz in identifying niche communities, exploring influencers’ networks, and uncovering potential collaborations. The results provide valuable design insights for future interactive tools focused on social media influence in the fashion industry.

The assistance of AI plays a significant role in non-native English speakers’ (NNES) formal writing. To investigate the impact of AI on cover letter writing, we designed a survey-embedded experiment. In this study, participants read and evaluated the hireability and writing quality of cover letters created by 4 types of authors respectively: NNES-edited, NNES-written, AI-edited, and AI-written. We ultimately received 118 valid responses from native English Speakers (NES). The results indicated that (1) AI-edited cover letters perform significantly better than other three types of cover letters, but (2) The editing of NNES made the performance of AI-written cover letter worse. Our research provides insights into how artificial intelligence (AI) impacts NNESs with regards to cover letter writing, a critical part of the job application process. This outcome allows us to conclude that NNES would greatly benefit from using AI as an assistive writing tool during their job search.

Community social service organizations are indispensable for supporting vulnerable populations with multifaceted challenges, such as housing instability and food insecurity, particularly in high-poverty areas where access to essential resources is limited. Effective information and knowledge sharing enables these organizations to offer timely and comprehensive services to these groups. Despite increasing research on social services and emerging technologies, little is known about how knowledge is shared in community-based social service settings or how technology can enhance these complex processes. This paper presents findings from an ethnographic study conducted at four community organizations in an underserved community in the Midwest of the United States. Through observations and thematic analysis, we identify key stakeholders that participate in knowledge sharing, including staff and service users, and examine their roles in developing Communities of Practice (CoP). Our study reveals the dynamics of knowledge sharing and highlights opportunities for leveraging technology to support CoP within community social services. These findings contribute to discussions on designing socio-technical systems for improving information sharing in resource-constrained environments.

This extrapolation explores the evolution of Personal Knowledge Management (PKM) and envisions its future integration with artificial intelligence (AI). As we transition from traditional organizational systems to sophisticated digital ecosystems, the concept of a ’second brain’ has emerged, exemplified by tools like Evernote and Notion. However, the integration of AI promises to transform this concept into an active personal companion. This AI-driven system would access multiple data streams, creating a rich, interconnected knowledge base that offers personalized insights and decision support. The paper discusses the potential design of such an AI companion. Unlike current tools that excel at organizing information, this AI companion would actively engage with data from various aspects of a user’s life, creating a dynamic, personalized knowledge overview. While the potential benefits are significant, the paper also addresses critical considerations, including privacy concerns, ethical implications, skill requirements for effective use, and the need to balance human intuition with machine intelligence. The discussion emphasizes the importance of maintaining user autonomy and critical thinking skills while leveraging AI capabilities. As we enter "The Intelligence Age," this extrapolation provides a foundation for further research and discussion on the responsible development and implementation of AI companions as advanced cognitive tools, aiming to augment human intelligence rather than replace it.

Browsing social media feeds in bed can reduce relationship quality, disrupt sleep, increase anxiety, and cause physical discomfort. To provoke creative discussion around these issues, this paper presents a demonstration of a refreshing alternative. “Weight of Scrolling” is an interactive installation that successfully reimagines social media scrolling through a tangible embodied media interface that fosters positive collaborative engagement. Participants are invited to lay down together and collectively, physically manipulate a large-scale cotton fabric scroll of simulated social media updates.

In emergency medicine, where delays can result in dire consequences, technological interventions have the potential to significantly improve outcomes. This research focuses on understanding trauma resuscitation teams’ experiences with delays and designing an alerting system to improve delay awareness during time-critical teamwork. This research has three aims: (1) identify critical delays and examine delay awareness, (2) define design requirements for delay alert systems, and (3) evaluate the impact of visual alerts on delay awareness. Through iterative user research, we developed a prototype alert system. Current and future work involves designing an ambient alert system and evaluating its effects through live simulations.

In June 2023, the U.S. Supreme Court banned race-based affirmative action (AA) in college admissions, reshaping college access for underrepresented students. AA is historically controversial, and social media platforms enable discussions of AA that traverse physical and social barriers. My dissertation explores these social media discussions of AA, taking a qualitative, cross-platform approach to investigate how socio-technical properties of platforms shape social media-based AA discussions and their impacts, especially for underrepresented college applicants. I aim to make empirical, theoretical, and design contributions to promote informed, mindful, and deliberative discussions of contentious, identity-related topics on social media.

My doctoral research examines the intersection of the gig economy and self-tracking, exploring how gig workers use self-tracking to navigate their jobs. It outlines two main objectives: first, to investigate gig workers’ self-tracking behaviors and practices through interviews and surveys; and second, to design effective tools for gig workers by systematically evaluating self-tracking apps and enhancing their design through co-design workshops. By integrating findings from these studies, the dissertation aims to enhance self-tracking systems, improve gig worker support, and contribute to human-computer interaction (HCI) research.

The massive datasets used to train AI models frequently contain offensive and harmful entries that are only caught during the system’s later performance. The data workers who curate such datasets are experts in datasets’ contents, but silenced by horrible labor conditions and lack of respect by their employers. In my dissertation, I study how to build 1) workplaces, 2) workflows, and 3) tools to elicit and embrace data workers’ observations in dataset development. My goals are creating datasets safe to use to train AI systems and developing a more pro-social data labor paradigm.

100% of us will die. Meanwhile, each of us creates more and more online accounts and data throughout our lives than ever before. Although online accounts and data must increasingly be integrated with our end-of-life planning practices, that integration has been slow. Inventory-based data planning approaches recently developed by legal firms and industry players show promise in providing support. However, these approaches are still largely focused on planning for financial assets like online bank accounts. It is unclear if these approaches address the deeper human-centered needs, such as digital legacy. To imagine new approaches for end-of-life planning that account for emerging social-technical needs, this dissertation asks: How might social computing researchers and practitioners design more effective approaches for end-of-life data planning that can holistically account for human factors such as digital legacy?

Infrastructural limitation in different forms such as in technological, social, and political aspects affect the adoption and use of technology. Thus, it is important to understand how digital technology is appropriated for current and future use in such contexts. This research investigates how digital technologies are used and appropriated in environments characterized by infrastructural limitations, focusing on grassroots efforts in Bangladesh and the United States. Using qualitative methods and participatory approaches, the research aims to characterize existing socio-technical strategies employed by users to navigate infrastructural limitations, and to co-design future directions for interactive technology to leverage existing practices to overcome the limitations of infrastructure and power dynamics.

Developing AI literacy among students is vital in an AI-driven world. While scholars have explored AI literacy in the CS context, some critical areas remain under-examined for technology students who study disciplines such as cybersecurity, information systems, and information technology, where they are often expected to apply AI tools and techniques rather than develop them from scratch. In this research leading toward my dissertation, I focus on three under-examined areas: 1) the effect of collaborative learning on AI literacy development; 2) the role of students’ group engagements with everyday AI; and 3) students’ development of an AI-ethics mindset.

Generative AI (GAI) systems are increasingly integrated into children’s digital experiences, raising concerns about safety, privacy, and appropriate interactions. My research explores the specific risks children encounter when interacting with GAI, with a focus on identifying harmful interactions and their potential impacts. By examining parent-child perceptions of GAI safety, I aim to develop a comprehensive framework and database of risky interactions to guide the creation of effective parental control tools and educational resources. The ultimate goal is to provide parents with practical solutions to actively mitigate risks and create safer AI environments for children.

This research develops scalable and interpretable learning tools—stealth assessment and a teacher dashboard—using a sophisticated logging system within the Mission HydroSci game-based learning environment. By integrating existing frameworks with customized approaches, the system captures in-game behaviors critical to understanding learning processes. A stealth assessment pipeline, combined with advanced machine learning, and a data-enriched dashboard, is being developed to enhance teacher-student communication. This study contributes a scalable framework for learning tools across diverse game contexts, offering valuable insights for educators, researchers, and game designers to improve both learning and educational game design in classroom settings.

This dissertation examines how decentered users appropriate technologies amid rapid technological changes, focusing on digital activists, small business owners, and novice artists. It establishes a taxonomy for technological change for HCI research, emphasizing user interactions with platform technologies and concepts like appropriation and infrastructuring. The study also highlights how technological change can marginalize users by imposing rigid structures that often lack community-centered approaches, undermining infrastructures that support their needs. Ultimately, the dissertation hopes to propose design, policy, and theoretical contributions to create more inclusive systems that better serve the needs of decentered users.

Technology is pervasive in everyday family life. However, creating meaningful sociotechnical systems that are useful for (and desired by) families remains a complex and evolving challenge. To design for families in the digital age, HCI research requires holistic understanding and in-depth investigation to unpack social dynamics and situated use beyond the individual. My dissertation applies a family-centered approach to examine how technologies might be better designed to support everyday social, health, and developmental goals and challenges. Focusing on family-food interactions, my work unpacks complexities in navigating various food-related experiences to inform design implications.

Interactive content in online spaces is often meant to inform a broad audience regarding issues of societal interest. Recently, combating misinformation has emerged as another task requiring interventions to reach broad audiences. We are engaged in two efforts to investigate how lay users leverage social media to inform and engage a broader audience and the characteristics of the resulting user-generated discourse. We found that tailoring content for broader audiences can be laborious and the impact of the efforts is neither guaranteed nor immediately apparent. Our future work will focus on approaches to help lay users mitigate these problems.